VOL
7, NO. 1 1 , N O V E M B E R 1 9 6 8
An Enzyme System for de Novo Biosynthesis of Glycogen in Aerobacter aerogenes* Linda C. Gahant and H. E. Conrad
ABSTRACT: A glycogen synthetase fraction and an activator protein, which together catalyze de novo glycogen synthesis from adenosine diphosphate glucose, have been purified 400- and 40-fold, respectively, from extracts of late log cells of Aerobacter aerogenes A3(S1) and have been shown to be devoid of glycogen. In the presence of adenosine diphosphate glucose and magnesium chloride the purified glycogen synthetase catalyzes a slow rate of glycogen synthesis which is stimulated several hundredfold by supernatant activator protein. Similar stimulations of the glycogen synthetase activity are obtained with glycogen, bovine serum albumin, Triton X-100, several A . aerogenes lipopolysaccharides, and the lipid A fraction of the A3(S1) lipopolysaccharide. The diversity of physical and chemical properties of the compounds which stimulate glycogen synthetasesuggests that theyare all activatorsof glycogen synthetase rather than acceptors for the glucosyl resi-
G
lycogen synthetases, regardless of their source, have shown an absolute requirement for a glycogentype glucan to accept glucose from the system-specific nucleoside diphosphate glucose (Leloir and Cardini, 1962; Greenberg and Preiss, 1965; Preiss and Greenberg, 1965). In previous work (Kindt and Conrad, 1967) it was shown that extracts of late log cells of Aerobacter aerogenes A3(S1) would synthesize glycogen from ADP-glucose only when exogenous glycogen was added, in spite of the fact that up to 2 g of glycogen/100 g of dry cells could be extracted from stationary-phase cultures of this organism. The tentative conclusion based on these observations was that late-log cells, in contrast to stationary-phase cells, were devoid of glycogen. This has been demonstrated to be the case (Koeltzow et ai., 1968) but the question has remained, “How is glycogen synthesis activated when the stationary growth phase begins ?” In the earlier work several unique features of the A. aerogenes glycogen synthetase were reported. Unfractionated extracts and a particulate preparation (fourfold purified) obtained from the crude extract by centrifugation at 105,OOOg were studied. Both preparations
* From the Division of Biochemistry, University of Illinois, Urbana, Illinois 61801. Received June 27, 1968. Supported by a grant from the U. S . Public Health Service, National Institute of Allergy and Infectious Diseases (AI 05696). t Holder of a National Institutes of Health predoctoral traineeship on U.S . PublicHealth ServiceTraining Grant GM-321.
dues transferred from adenosine diphosphate glucose. As glycogen synthetase loses activity on standing it can be activated only by glycogen and by bovine serum albumin and thus becomes more like a typical glycogendependent glycogen synthetase. The mechanism of the de novo synthesis has been studied by pulse-labeling experiments. These show that, of the glucose residues incorporated into glycogen in the early stages of the reaction, approximately 50% remain in the external chains of the product recovered after a four- to sixfold increase in the amount of glycogen synthesized. This observation indicates that the de noco process differs from the usual mechanism for glycogen biosynthesis and suggests a mechanism like that found for 0-antigen biosynthesis (Robbins, P. W., Bray, D., Dankert, M., and Wright, A. (1967), Science 158, 1536) in which repeating oligosaccharide units are polymerized from the reducing end of the growing polysaccharide chain.
catalyzed glucosylation of less than half of the primer nonreducing terminals and formed highly branched glycogens. Apparent primer K M values for unfractionated cell extracts, expressed as concentration of nonreducing terminals, showed a progressive increase with decreasing chain length of the primer. Upon pelleting the activity the primer K M values were increased three- to eightfold and the correlation of K Mwith chain length of primer was lost. The kinetic parameters observed for initial cell extracts were restored upon recombination of the 105,OOOg supernatant and pellet. These observations implicated at least two fractions in the glycogen-synthesizing system of A . aerogenes, with the activities for glucosyl transfer and branch formation both in the pellet fraction. The purification and partial characterization of these two fractions are described here. The glucosyl transferase and branching activities are recovered together in a 400-fold-purified GS1 fraction. This fraction actively synthesizes glycogen only when SAP is added. Evidence ~~
The abbreviations used that are not given in Biochemistry 5 , 1445 (1966), are: BSA, bovine serum albumin; GS, glycogen synthetase; SAP, supernatant activator protein; DTT, dithiothreitol; ECx,the average number of glucose residues in an external chain of glycogen; ICL, the average number of glucose the average number residues in an internal chain of glycogen; of glucoses in a chain length of glycogen (TL = F L E L 1); LPS, lipopolysaccharide; PS, the polysaccharide portion of a lipopolysaccharide; lipid A, the lipid portion of a lipopolysaccharide. 1
a,
+
de Noco GLYCOGEN BIOSYNTHESIS
IN
Aerobncter
+
3979
aerogenes
BIOCHEMISTRY
is presented to show that neither of the purified fractions contains glycogen. The supernatant fraction can be replaced by BSA, lipopolysaccharide, lipid A, Triton X-100, or glycogen. In all cases a highly branched product is formed. Pulsing experiments show that approximately 50% of the glucose residues incorporated into glycogen in the early stages of synthesis remain in the nonreducing periphery of the molecule as glycogen synthesis proceeds. The data suggest that the de nouo process may proceed by a mechanism similar to that described for 0-antigen biosynthesis (Robbins et ai., 1967; Osborn and Weiner, 1968) rather than by the classical glucosyl transferase-branching enzyme mechanism. Methods
3980
[ 4CIADP-glucose was obtained from New England Nuclear Corp. and diluted with unlabeled ADP-glucose (Calbiochem) to 0.1 mCi/mmole for use in the enzyme assay. Crystalline sweet potato P-amylase (400 units/mg of protein) was obtained from Worthington Biochemical Corp. and crystalline swine pancreas a-amylase (500 units/mg of protein) from Sigma Chemical Co. One unit of these enzymes, assayed as described by Bernfeld (1955), releases from amylopectin 1 pmole of maltose equiv/min at 25 ". Crystalline bovine pancreas trypsin (40 units/mg of protein; 1 unit hydrolyzes 1 pmole of p-tosyl-L-arginine methyl ester/min at 25") was obtained from Calbiochem and assayed by the method of Humme1 (1959). Crystalline soybean trypsin inhibitor and crystalline pancreatic ribonuclease were purchased from Worthington Biochemical Corp. ; DTT, Triton X-100, and p-tosyl-L-arginine methyl ester from Calbiochem; and crystalline pancreatic deoxyribonuclease from Sigma Chemical Co. Paper chromatograms were run in 1-butanol-pyridine-water (6 :4:3) and standards were detected by spraying with aniline acid phthalate reagent (Partridge, 1949). Radioactive compounds on chromatograms were quantitated by cutting the chromatograms into segments and counting the segments as before (Kindt and Conrad, 1967). The glycogen used throughout this work was isolated from stationary-phase A . aerogenes A3(S1) cells as for the previous work (Kindt and Conrad, 1967). LPS from several strains of A . aerogenes and the lipid A and polysaccharide fractions from the A3(S1) LPS were prepared as described by Koeltzow et al. (1968). The characterization of the biosynthetic glycogen by methylation analysis and &amylolysis and the use of chain length parameters to define the glycogen structure have been described (Kindt and Conrad, 1967). Assay of GS. Assays were carried out at room temperature in 5 X 15 mm test tubes containing, per 50 pl: Tris-HC1 (pH 7.4), 2 pmoles; magnesium chloride, 1.25 prnoles; [ 14C]ADP-glucose, 0.25 pmole (25,000 cpm when counted as below); BSA, 100 pg; glycogen, 25 pg; and enzyme. Reaction was initiated by adding GS (0.05-0.4 unit) or, in some cases, ADP-glucose. At 7.5, 15, 30, and 45 min 10-pl aliquots were removed with a n Eppendorf pipet (Brinkmann Instru-
G A H A lrl A N D
CONRAD
ments, Inc.) and spotted on a strip of Whatman No. 4 paper, and the strip was dipped into acetone to stop the reaction. The paper strip was 1 in. wide and 3 in. long and was marked off in 0.5-in. segments numbered 1-6. Successive aliquots from the assay were streaked on segments 2-5. Segment 6 was cut to a drip tip and segment 1 was attached to a thick paper wick (four 1 X 2 in. segments of Whatman No. 3MM paper stapled together) with a paper clip. The unreacted substrate was eluted from the paper strip with 50% ethanol in 0.1% aqueous ammonium hydroxide for 1 hr, then rinsed with acetone and air dried. Segments 2-5 were cut out of the strip and counted to &3% error in a liquid scintillation counter as before (Kindt and Conrad, 1967). Rates of glycogen synthesis are calculated from the counting data as micromoles of glucose incorporated into the 50% ethanol-insoluble product. A unit of GS catalyzes incorporation of 1 pmole of glucose into glycogen per hr under these assay conditions. Assay of SAP. SAP was assayed as above, using a reaction mixture containing, per 50 pl: Tris-HCI (pH 7.4), 2 pmoles; magnesium chloride, 1.25 pmoles; [14C]ADP-glucose, 0.25 pmole (25,000 cpm); BSA, 100 pg; 176,OOOg pellet fraction containing GS activity (step 2, Table 11), 50 pg; and SAP, 0.05-0.25 unit. One unit of SAP induces an increase in the rate of glycogen synthesis of l pmole of glucose incorporation/hr over that obtained with the same assay system in the absence of SAP. Protein was determined by the method of Lowry et al. (1951). Specific activities are expressed as units per milligram of protein. Purification of Components of the Glycogen-Synthesizing System. A . aerogenes A3(S1) (ATCC 12658) was grown at 30" on a glucose-mineral salts medium, harvested in the late logarithmic growth phase, and washed as before (Kindt and Conrad, 1967). Frozen cell paste (14 g) was broken in a Hughes press and suspended in 34 ml of 0.05 M Tris-HC1 buffer (pH 7.4) containing 0.001 M DTT (Tris-DTT buffer). All subsequent operations were carried out at 0-5 ". Deoxyribonuclease (0.5 mg) and ribonuclease (0.5 mg) were added to depolymerize the nucleic acids. After 45 min at 0" the liquified mixture was centrifuged at 32,OOOg for 15 rnin and the pellet was washed with 20 ml of Tris-DTT buffer and recentrifuged. The combined supernatants were centrifuged for 150 min at 176,OOOg in a 65 rotor of a Spinco L2-65 ultracentrifuge to pellet most of the GS activity and leave the SAP activity in the 176,OOOg supernatant. The supernatant was used for further purification of the SAP; GS activity was purified from the pellet fraction. Further Purification of SAP. To 54 ml of the 176,OOOg supernatant 10.6 g of ammonium sulfate was added with stirring to give 33% saturation. After 20 min the resulting precipitate was collected by centrifugation at 32,OOOg for 10 min, suspended in 4 ml of Tris-DTT buffer, and dialyzed against this buffer overnight. The dialyzed fraction was chromatographed on a 1.3 x 10 cm DEAE-cellulose column equilibrated with Tris-DT'T buffer, using as eluate a linear gradient made up of 110 ml each of the Tris-DTT buffer and 0.5 M sodium chloride containing 0.001 M DTT. The column was de-
VOL.
7,
TABLE I:
NO.
11,
NOVEMBER
1968
Purification of Supernatant Activator Protein.. Activity Fraction
Vol (ml)
Protein (mg)
Unitsb
Recov
(%)
Sp Act. (units/mg)
1. 176, OOOg supernatant 2. AmS04 precipitate, 0-33% saturation 3. DEAE-cellulose eluate, 0.20-0.28 M NaCl 4. Concentrate of DEAE-cellulose eluate
54.0 5.9 39.6
836.0 82.5 17.3
4010 3080 2930
100 77 73
4.8 37.4 170.0
8.0
16.2
2900
72
179.0 ~~
~
= Starting material was 14 g of cell paste. b One unit of SAP induces the incorporation of 1 pmole of D-glucose/
hr from ADP-glucose into glycogen in an assay system containing, per 50 pl: Tris-HC1 (pH 7.41, 2 pmoles; MgClz, 1.25 pmoles; [14C]ADP-glucose,0.25 pmole(25,OOO cpm); BSA, 100 pg; GS, 0.3 unit; and SAP, 0.05-0.25 unit.
veloped at a flow rate of 0.15 ml/min and 3-ml fractions were collected and assayed for SAP. To concentrate the SAP the combined peak fractions of activity, emerging between 0.20 and 0.28 M sodium chloride, were dialyzed for 6 hr against Tris-DTT buffer and adsorbed on a 1.3 X 5 cm DEAE-cellulose column .equilibrated with Tris-DTT buffer. The activity was washed directly through the column with 30 ml of 0.4 M sodium chloride containing 0.001 M DTT and stored at 4" in the elution solvent. Further Purification of Glycogen Synthetase. The 176,OOOg pellet from above was allowed to stand 14 days at 4" and was then suspended in 13 ml of cold saline and added to 133 ml of 0.025 M Tris-HC1 buffer (pH 7.4) containing 0.001 M DTT and 0.23 ml of Triton X-100. After standing for 1 hr at 0" the solution was centrifuged at 100,OOOg for 1 hr and the supernatant, containing the solubilized GS, was separated carefully with a pipet from the pellet. To the solubilized protein was added 24 g of solid ammonium sulfate (to give 30% saturation) and, after stirring for 20 min, the suspension was centrifuged at 32,OOOg for 10 min to bring the precipitated protein to the surface of the centrifugate. The underlying solution was carefully withdrawn with a Pasteur pipet and the precipitate was dissolved from the sides of the centrifuge tubes in 8 ml of 0.025 M Tris-HC1 (pH 7.4) containing 0.001 M DTT, dialyzed overnight against the same buffer, and applied to a 1.3 X 10 cm DEAE-cellulose column equilibrated with the above buffer containing 0.15% (w/v) Triton X-100 (TrisDTT-Triton buffer). The column was eluted with a linear gradient consisting of 120 ml each of Tris-DTTTriton buffer and 1 M sodium chloride containing 0.001 M DTT and 0.15% Triton X-100. The column was eluted at a flow rate of 0.5 ml/min and 3.5-ml fractions were collected and assayed for GS and protein. The active fractions, recovered between 0.35 and 0.47 M sodium chloride, were either stored at 4" in the eluting buffer or concentrated by dialyzing the pooled fractions against Tris-DTT-Triton buffer, adsorbing the protein on a 1.3 X 5 cm DEAE-cellulose column in the same buffer, and washing the activity from the column with 30 ml of 0.5 M sodium chloride containing 0.001 M Dm and
de
N O V O
0.15% Triton X-100. The concentrated fractions were pooled and stored at 4" with gradual loss of activity over a 2-week period. Results General Description of Glycogen-Synthesizing System. The glycogen-synthesizing system of A . aerogenes consists of two fractions which are separable by ultracentrifugation at 176,OOOg for 2.5 hr. The GS recovered in the pellet contains all of the catalysts necessary to synthesize a highly branched glycogen from ADPglucose but gives very low rates of glycogen synthesis unless activated by the supernatant fraction. This requirement for SAP may be replaced in part by BSA, but the BSA-activated GS is stimulated to still higher rates of glycogen synthesis by SAP, Le., the BSA and the SAP activation of GS are more or less additive. For assay during purification the SAP activity is determined by measuring the extent to which it will stimulate the activity of GS in the presence of BSA. BSA is included in the SAP assay since in its absence mixtures of the crude GS and SAP from the ultracentrifugation give nonlinear rates of glycogen synthesis. When the two fractions are further purified, however, they will give linear rates of glycogen synthesis when assayed together. Maximum activation of GS is obtained with a mixture of BSA and glycogen or with a mixture of BSA and SAP. Since the former activation mixture is made up of the most stable and reproducible components, it is used in the assay for purification of GS. In the results presented below, units of GS activity refer to that activity obtained with both BSA and glycogen in the GS assay. Where rates are given, they refer to the rate observed in the experiment described. Purification of SAP. The purification of SAP is shown in Table I. Initial activities are measured in the 176,OOOg supernatant which contains most of the SAP activity and approximately two-thirds of the protein present in cell extracts. A 37-fold purification is obtained with a 72% recovery of units. In development of the DEAEcellulose chromatography step it was observed that more shallow salt gradients than used in the purification
GLYCOGEN BJOSYNTHESIS
IN
398 1
Aerobacter aerogenes
BIOCHEMISTRY
TABLE 11:
Purification of Glycogen Synthetase System.a Activity Recov
Sp Act. (units/mg)
Fraction
Vol (ml)
Protein (mg)
Unitsb
(%)
1. Cell extract, 32, OOOg supernatant 2. Pellet, 176,000g 3. Triton-solubilized pellet, 100,OOOg supernatant 4. AmS04 precipitate, 0-33% saturation 5. DEAE-cellulose eluate, 0.35-0.47 M NaCl 6. Concentrate of DEAE-cellulose eluate
57.0 144.0 138.0
1175.0 429.0 236,O
3660 2400 2060
100 66 56
3.1 5.6 8.7
5.4 36.0
37.8 1.2
2000 1720
55 47
43.0 1390.0
10.8
1.4
1680
46
1200.0
0 Starting material was 14 g of cell paste. One unit catalyzes the incorporation of 1 pmole of D-ghcose/hr from ADP-glucose into glycogen in an assay system containing, per 50 pl: Tris-HCI (pH 7.4), 2 pmoles; MgC12, 1.25 pmoles; [14C]ADP-glucose,0.25 pmole(25,OOO cpm); BSA, 100 pg; glycogen, 25 pg; and GS, 0.05-0.40 unit.
procedure finally adopted gave a marked smearing of the SAP activity over a number of column fractions. The concentrated SAP is devoid of GS activity and can be stored for at least 6 months at 4" without loss of activity. Purification of GS. The purification of GS is shown in Table 11. For fractionation of the 176,OOOg pellet, the GS is solubilized by extraction of the pellet with 0.15% Triton X-100 in 0.05 M Tris-HC1 buffer (pH 7.4). It was found that this concentration of detergent extracts 100% of the GS activity, provided the pelleted protein is allowed to age for 2 weeks at 4". GS activity is slowly lost after longer periods of aging but is incompletely extracted when the pellet is aged less than 2 weeks. Detergents other than Triton X-100 (e.g., sodium dodecyl sulfate, sodium cholate) either are less effective in extraction of activity or inhibit glycogen synthesis. Triton X-100 is needed not only to solubilize the activity but to keep it in solution. Dialyzed solutions of
TABLE ITI: Requirements for Glycogen Synthesis.
Assay System
Glucose Incorp (pmole/hr)
Complete. -GS -SAP - BSA -BSA and SAP -SAP, glycogen; 25 pg -MgC12
0.34 0.00 0.25 0.27 0.00 0.34 0.05
+
3982 GAHAN
a The complete system contained, in 50 pl: Tris-HC1 (pH 7.4), 2 pmoles; MgCI2, 1.25 pmoles; [14C]ADPglucose, 0.25 pmole (25,000 cpm); BSA, 100 pg; SAP, 2.25 units (259 units/mg, see Table I); and 0.34 unit of GS (680 units/mg).
AND CONRAD
solubilized enzyme become cloudy and lose activity rapidly.2 When detergent is not included in the buffer used for elution in the DEAE-cellulose purification step, much of the protein and activity precipitates at the top of the column and only 15-20Z of the activity is recovered and is smeared over a number of fractions. In contrast, when the buffer containing Triton X-100 is used, the activity emerges from the column in a very sharp peak which follows most of the protein. The purification gives a 400-fold over-all increase in specific activity and yields a product which loses its activity rapidly on freezing and thawing but more slowly when stored at 4". Requirements for Glycogen Synthesis. The data in Table I11 show that glycogen synthesis by the purified system requires GS, MgCl9,and an activator or a combination of activators. As indicated above both SAP and BSA serve as activators but alone neither one induces the maximum activity of the GS. Table I11 shows that glycogen plus BSA will replace SAP plus BSA in inducing the maximal activity of GS. It is because of this maximum stimulation obtained with BSA plus glycogen that both activators are used in the assay system for GS purification. The Actiaity of GS in the Absence of Actieators. The rates of glycogen synthesis when increasing levels of purified GS are assayed in the absence of activator are shown in Figure 1. At the usual levels of GS which give high linear rates when activators are present (the three lowest levels in Figure 1) glycogen synthesis proceeds only after a lag period which is gradually eliminated as the level of GS is increased. Only at high levels of GS 2Triton X-100 has a molecular weight of 635 and exists in aqueous solution as a micelle with approximately 100 molecules/micelle (Dwiggins et al., 1960). The micelle-monomer equilibrium permits the dialysis of the detergent. Its loss from aqueous solution on dialysis has been demonstrated by a decrease both in absorbancy of the solution at 280 mp and in the weight of detergent recoverable.
VOL.
7,
NO.
11,
TABLE IV:Activation
NOVEMBER
1968
of Glycogen Synthetase.a
Activator None SAPb Glycogen BSA Triton X-100 Boiled SAP Boiled glycogen Boiled GS Boiled BSA Lipopolysaccharidesc NCTC 9652 NCTC 9656 NCTC 9644 NCTC 9504 NCTC 418 NCTC 243 A3(S1) A3(S1) PS A3(S1) lipid A
b
.s} a
Amt (Hg/ 50 ~ 1 )
Glucose Incorp (Wmoleihr)
4.5 5 .O 6.0 50.0 8.7 5.0 0.5 6.0
0.00 0.23 0.22 0.16 0.17 0.00 0.22 0.00 0.00
.2
GS
HOURS
25.0 25.0 25.0 25.0 25 .O 25.0 25.0 200.0 100.0
0.00 0.00 0.00 0.06 0.15 0.15 0.13 0.00 0.14
.6
protein
.8
Glycogen synthesis catalyzed by purified GS. Assays were performed as in Table 11, but without glycogen or BSA. Specific activity of GS (step 5, Table II), 715 units/ mg. (a) The lag in glycogen synthesis as a function of GS level in assay. (b) Linearity with GS of maximum rate attained after lag. FIGURE 1:
a Assay with 0.3 unit of GS (505 units/mg) as described in Table 11,but without BSA and glycogen except where indicated. b Specific activity of SAP was 259 units/mg. c All lipopolysaccharides tested were from A . aerogenes strains.
does glycogen synthesis proceed without a lag. As shown in Figure lb, the maximum rate of synthesis attained after the lag is a linear function of the amount of GS in the assay. These maximum rates are only onethird of the rates obtained in the presence of BSA and glycogen. In the 45-min period used in the normal assay the three lowest levels of GS shown in Figure l a would be considered completely inactive in the absence of activators. Activity of Glycogen Synthetase in the Presence of Actioators. A number of substances, when added to the assay system with GS, eliminate the lag at the low GS levels and increase the maximum rate of glucose incorporation attainable at all G S levels. The relative effects of a variety of activators on the rate of glycogen synthesis are shown in Table IV. All of these activators induce linear rates of glycogen synthesis from the time of initiation of the assay. In addition to the activation by SAP, BSA, and glycogen described above, Triton X-100 and the lipopolysaccharides from some (but not all) strains of A . aerogenes activate the GS. Since the 45-min assay period was used in these experiments any onset of activity that might have occurred after a lag period was not recorded. Several points merit special comment. The effect of glycogen in the assay is identical with that of the nonglycogen activators. The possibility therefore arises that glycogen functions primarily as an activator rather than as a glucosyl acceptor. The fact
de N o a o
4
- up
that when SAP, GS, or BSA are boiled they do not serve as activators is evidence that these preparations do not contain significant levels of glycogen though it is possible that the denaturation-aggregation process renders any glycogen that may be present inaccessible to the enzyme system. A second observation of note is that when the A3(S1) LPS is separated into its PS and lipid A components, the activator activity is found only in the lipid A fraction. It is because of the diversity of physical and chemical properties among the group of substances which stimulate GS that these compounds are considered to be activators rather than glucosyl acceptors. Further evidence to indicate that SAP is not a glucosyl acceptor is that upon ultracentrifugation at 176,OOOg of an assay mixture in which SAP was used to activate [l?C]glycogen synthesis, none of the I C in the product is recovered in the supernatant with SAP. Figure 2 shows the linear relationship between the rate of glycogen synthesis and the amount of GS added to assay systems containing saturating levels of activators or combinations of activators. When no activator is present only the highest levels of GS assayed give measurable initial rates of glycogen synthesis (cf. Figure 1). In the presence of activators the rates of glycogen synthesis are directly proportional to the amount of GS in the assay. The partially additive activation effects of BSA and glycogen or BSA and SAP and the lack of additive effects with glycogen and SAP suggest that glycogen and SAP act at the same site on GS while BSA acts at a different site. In experiments not shown here it has been found that when glycogen and SAP are both present at subsaturating levels in the assay their activation effects are additive. The saturation of GS with various activators is shown in Figure 3. On a weight basis, glycogen and SAP are most active. Since no attempt has been made to bring SAP to a high level of purity for these studies, it is anticipated that, on a weight basis, SAP will prove to be
GI,YCOGF.N
BIOSYNTHESIS
IN
3983
Aerobacter aerogenes
BIOCHEMISTRY
TABLE v:
Change in G S Activity with Different Activators during St0rage.a G S Activity (% of original) Conditions of Storage
SAP
Triton X-100
Lipid A
BSA
Glycogen
Freshly prepared, step 5, Table I1 3 days, 4" 7 days, 4" 15 days, 4" 20 hr, -10" (frozen)
100 100 41 2 0
100 100 55 2 33
100 95 41 2 36
100 95 82 20 86
100 98 95 20 90
a Purified GS (1510 units/mg) stored either at 4" or at -10". The preparation stored at 4" was in the buffer used to elute the activity from DEAE-cellulose (step 5, Table 11). The preparation stored at -10" had been concentrated as in step 6, Table 11, and then made 50% saturated with ammonium sulfate. All assays contained 0.5 pg of GS. Activator levels used were: 6.8 pg of SAP (259 units/mg); 50 pg of Triton X-100, lipid A, or BSA; and 25 pg of glycogen.
much superior to glycogen as an activator for GS. No data have been obtained concerning the molar concentrations for optimum activator activity. Effect of Aging of GS on Activation. When the purified GS stands at 4" or is frozen at -10" with ammonium sulfate in the buffer used to elute it from DEAE-cellulose, it loses activity and the relative activities of different activators change as shown in Table V. Most significant is the fact that the activity obtained
with SAP, the natural activator of the system, is lost quite rapidly while that with glycogen and with BSA is retained for the longest period. Thus, the de nor0 system for glycogen biosynthesis reverts to a system similar to those previously described (Leloir and Cardini, 1962; Greenberg and Preiss, 1965; Preiss and Greenberg, 1965) in which glycogen is required in the assay and BSA stimulates the rate of glycogen synthesis. In the earlier work with this system (Kindt and Conrad, 1967) this reversion apparently took place when cell extracts were prepared by sonication, since such ex-
NO ACTIVATOR
0.2
0.4
0.6
GS-
3984
0.0
1.0
1.2
1.4
FIGURE 2: Linearity of initial rates of glycogen synthesis with GS level when saturating levels of activators are present alone, or in combination. Per 50 pl of assay volume the activator levels used were: BSA, 100 pg; SAP (259 units/ mg), 8.7 p g ; and glycogen, 25 pg. Rates of glycogen synthesis were linear over the 45-min assay period except at the low GS levels when no activator was present.
G A H A N AND CONRAD
ACTIVATOR- ug /50ul
yg prot@in
3: Saturation of GS with activators. Assays were performed with 0.30 unit of GS (507 units/mg) as in Table I1 but with replacement of the BSA and glycogen with the various activators at the levels indicated. Enzyme was dialyzed us. Tris-DTT buffer prior to assay. For SAP the amounts added are given in micrograms of protein; for other activators amounts are in micrograms dry weight.
FIGURE
VOL.
7,
TABLE VI:
11,
NO.
NOVEMBER
1968
Properties of Supernatant Activator Protein. Glucose Incorp (pmoles/hr)
Additions to Assay. Experiment 1 a. None b. SAP and a-amylase
}
4.50
I
0.26 0.37
c* SAP and preincubated" a-amylase ~. d. Glycogen and a-amylase
,y
z
-0!a
0.35
8
0.50
0.10
a a
0 0
-z Experiment 2. a. Trypsin and trypsin inhibitor b. SAP, trypsin, and trypsin inhibitor c. SAP and trypsin, preincubated trypsin inhibitor
w v)
0
0.52 1.44
2
0.05
J
c3
0.49
5 Assays as described in Table 11, but without glycogen and BSA, except where indicated. The GS preparation used had a specific activity of 411 units/mg and, due to the Triton X-100 in the buffer in which it was stored, gave a measurable rate of glycogen synthesis in the absence of added glycogen or SAP. For experiment 1, 0.76 unit of GS was used; for expt 2, 1.51 units. When added, 25 pg of glycogen or 2.25 units of SAP (259 units/mg) was present in the assays. * The indicated assay mixtures were preincubated with 0.0005 unit of a-amylase for 2 hr at room temperature. The ADPglucose and GS were omitted from the preincubations and were added simultaneously to the assay mixtures to initiate the assays. .Trypsin and trypsin inhibitor were added to the assays in the following manner. A 2.5-pl aliquot of a trypsin solution (5 mg/ml, 40 units/ mg) was mixed with 0.5 p1 of 1 M CaClz and 50 p1 of buffer (expt 2a) or a solution (step 3, Table I) of SAP (expt 2b,c). To 10 pl of this mixture was added 2.5 p1 of a solution of trypsin inhibitor (1 mg/ml, 40 units/ mg). After a 5-min incubation period 10 pl of the latter solution was added to the assay mixture and the assay was performed as usual. For the preincubation experiment (2c) trypsin and SAP were incubated for 18 hr at room temperature before adding trypsin inhibitor. Under the preincubation conditions SAP was completely stable when no trypsin was present.
tracts were almost completely dependent upon glycogen for activity (BSA was not tested as an activator). The effect of increasing concentrations of SAP on aged GS is shown in Figure 4. Whereas with freshly prepared GS 0.45 pg of SAP protein induces almost maximal GS activity (Figure 3), this amount of SAP gives very little activity with aged GS in the regular assay period but does induce the aged GS to synthesize glycogen after a 1-hr lag. The lag period is reduced and
de Nooo
I
3
2
HOURS FIGURE 4: Activation of aged GS with SAP. Assays were performed with an aged preparation of purified GS as described in Table 11, but with substitution of SAP (259 units/mg) for BSA and glycogen. The freshly prepared GS (1200 units/mg) was aged at 4" for 2 weeks in the eluting buffer used in the DEAE-cellulosechromatography step of purification (step 5, Table 11) and 0.31 unit was used in these assays. Final specific activity (with glycogen plus BSA), 330 units/mg.
finally eliminated as more SAP is added. The maximum activity attainable with SAP after GS is aged 2 weeks is about 2 0 z of its activity when freshly prepared. Properties of Supernatant Acticator Protein. The purified SAP is a heat-labile, nondialyzable protein which contains no glycogen. It was demonstrated previously (Koeltzow et ut., 1968) that A . aerogenes A3(S1) cells grown under the conditions used here contain less than 7 pg of glycogen/100 g of cell paste. Data in Table IV have shown that the purified SAP loses all activating activity when boiled under conditions where glycogen is completely stable. The absence of glycogen in purified SAP is confirmed by the data in expt 1 of Table VI which show that an a-amylase treatment, which destroys the activating activity of glycogen, has no effect on the activity of SAP. The experiment was performed by preincubation of the activators for 2 hr with a level of a-amylase sufficient to destroy glycogen but too low to interfere markedly with the assay. Experiment 2 in Table VI shows that preincubation of SAP with trypsin completely destroys its activating activity. In this experiment trypsin inhibitor, added at the beginning of each assay to prevent destruction of the GS by trypsin, had no effect on the glycogen synthesis rate. When SAP was preincubated in the absence of trypsin, it was completely stable. Absence of Glycogen in the Glycogen Synthetase.
GLYCOGEN BIOSYNTHESIS IN
3985
Aerobacter aerogencs
BIOCHEMISTRY
TABLE VII:
Activities of G S and SAP at Different Growth
Stages. la
GLY CO G E N
-a 0
E a.
in assay
0.20
I
z 0
5i0:
0
n
Growth Stage.
G Sb
SAP.
Early log Mid log Late log Stationary
0.36 2.58 7.04 5.91
3.60 3.31 3.65 4.17
.The A660 values at the time of harvest were: early log, 0.40; mid log, 1.00; late log, 2.25; stationary (3 hr after late log), 2.35. *Assay as described in Table 11, but without glycogen. Enzyme preparations assayed were at the step 2 stage of purification. assay as described in Table I. Enzyme preparations assayed were at the step 1 stage of purification.
8 0 z
Sp Act. (units/mg)
0.10
W v)
0
0
3 (3
G LY C 0G EN
1
Pylas' in assay
15
30
45
60
MINUTES FIGURE 5 : Effect of a-amylase on glycogen synthesis. Assays were run as in Figure 1, with 0.95 unit of GS (1500 units/mg) and no activator. Where indicated, a-amylase (500 units/ mg) was added at a level of 0.1 pg/50 rl. All assay components except [ 14C]ADP-glucosewere mixed and reactions were initiated by addition of [14C]ADP-glucose. For the analyses indicated by open symbols (n,O), a-amylase and ADP-glucose were added simultaneously. For the experiment indicated by closed circles, a-amylase was incubated with the assay mixture at room temperature for 15 min before initiation of reaction. Maltosaccharides were determined by spotting aliquots from the assays on 1-in. wide strips of Whatman No. 1 paper and chromatographing in 1-butanol-pyridine-water (6 :4: 3) for 40 hr to separate maltotriose and maltose from glycogen and unreacted ADP-glucose. The micromoles of glucose recovered in maltosaccharides was calculated from the total counts per minute in the maltose-maltotriose region of the chromatograms.
3986
The purified G S is not sufficiently stable at room temperature for the preincubation period required for the a-amylase destruction of glycogen as described above. Therefore, to test for the possible presence of glycogen in G S the following approach was used. If, in fact, glycogen is formed de nouo from ADP-glucose, then a-amylase, added t o the assay, should not interfere with the transfer of the first several glucose residues to the primary acceptor of the system but should cleave maltosaccharides from the growing glycogen molecule as it becomes large enough t o serve as an a-amylase substrate. Thus, in the presence of a-amylase the G S should catalyze the conversion of ADP-glucose into maltosaccharides. The data from such a n experiment are shown in Figure 5. The assay mixtures are prepared with all components but [14C]ADP-glucose, and aamylase is added to two assays but not to a third. Assays are started by the addition of ADP-glucose and the results obtained in the presence and absence of a-amylase are compared. Glycogen synthesis proceeds at a normal rate when no a-amylase is added, but in the
G A H A N
A N D
C O N R A D
presence of a-amylase no glycogen accumulates. The rate of incorporation of glucose into maltosaccharides, identified as maltose and maltotriose by paper chromatographic comparison with standards, approaches the rate of incorporation of [ 14C]glucose into glycogen by the system. The rates of maltosaccharide formation are identical whether ADP-glucose is added at the time of addition of a-amylase or when the assay mixtures is incubated 15 min with a-amylase before addition of ADP-glucose. The difference in rates of conversion of ADP-glucose into glycogen and t o maltosaccharides suggests that the incorporation of the first several glucose residues into a glycogen molecule may proceed somewhat slower than the rate of extension of the chains. When ADP-glucose is preincubated with a-amylase and then heated to destroy the a-amylase it serves as a substrate for glycogen synthesis and gives the same rate as unincubated ADP-glucose in an assay system containing purified G S and BSA. Thus, all components of the glycogen-synthesizing system from A . aerogenes have been shown t o be free of glycogen. Changes in Glycogen Synthesis Actioities during Growth. The specific activities of GS and SAP in cells at different stages of cell growth are shown in Table VII. The SAP activity is quite constant throughout growth, but the G S in late log cells shows a 20-fold increase in specific activity over that found for early log cells. Thus, the onset of glycogen synthesis and accumulation when the cells enter the stationary growth phase (Kindt and Conrad, 1967; Koeltzow et al., 1968) cannot be attributed t o a rapid rise in SAP at the end of the logarithmic growth phase. Instead, glycogen accumulation appears to be controlled at the level of ADPglucose pyrophosphorylase (Preiss et al., 1966) and glycogen synthetase levels (Govons et al., 1968). Structure of the Synthesized Glycogen. Structures of the products formed by the purified GS alone and by the GS when it is activated by BSA, Triton, glycogen, or SAP are all very similar to the structures obtained
VOL.
7,
TABLE VIII:
NO.
11,
NOVEMBER
1968
Direction of Glycogen Synthesis. Glucose in Glycogen (pmoles) Activators None BSA
ADP-glucose in Chase* 1F Unlabeled 4 c
Unlabeled Glycogen
4 c
Unlabeled
At Start of Chase
At Time of PAmylolysis
@-Amylolysisd (% of I4C)
0.026 0.026 0.067 0.067 0 068 0.068
0.140 0.150 0.333 0.244 0.486 0.330
42 44 49 54 46 27
a Assay as in Table 11, but without glycogen (25 pg) or BSA (100 pg) except where indicated. Assays contained 1.52 units of G S (760 units/mg). * Assays initiated with 0.25 prnole of [ 14C]ADP-glucose (600,000 cpm/pmole) and chased with 1.0 pmole of unlabeled ADP-glucose. The chase was added to each assay after 5 min in 2 pl of water and reaction was continued for 55 min. c Aliquots ( 5 p l ) were taken at intervals for assay and micromoles of glucose incorporated into glycogen were calculated taking into account the specific activity of the [14C]ADP-glucosein the particular assay and the dilution factor due to the aliquot added for the chase. d At the end of the period the remaining sample (42 pl) was adjusted to pH 4.8 with 2 p1 of 0.1 N HCl and 0.3 p1 of 1.6 M sodium acetate buffer (pH 4.8). A 10-pl aliquot was removed for measurement of the amount of 14Cin the glycogen in the usual way and 1 pl of P-amylase solution (1 mg/ml) was added to the remaining volume. After 30 min a IO-pI aliquot was taken to determine the amount of 14Cin the @-dextrin by the usual assay. Maltose, but not @-dextrin, is washed from the assay paper with the 50% ethanol-0.1 % NH40H.
with the initial cell extracts (Kindt and Conrad, 1967). The 14Cin these products is converted quantitatively to glucose by acid hydrolysis and to maltosaccharides by a-amylase. @-Amylase solubilizes 50-60% of the glycogen and yields maltose (identified by RF).The average chain length, determined by methylation analysis, was 10, somewhat lower than found previously for the system before purification. Direction of Glycogen Synthesis. Since the glycogensynthesizing system described here catalyzes glycogen formation in the absence of the nonreducing terminals of a primer, the direction of growth of glycogen during synthesis can be tested readily by pulse-labeling experiments. According to the current view of glycogen synthesis, illustrated in Figure 6, external chains of glycogen are extended at the nonreducing terminals by successive additions of glucose until they become long enough to serve as substrates for branching enzyme. The action of branching enzyme is to transfer an outer segment of the glycogen molecule from its position in the external chain to a C-6 hydroxyl on a residue in the external chain close to the previous branch point. Those glucoses added early in the reaction should become buried further and further in the internal part of the molecule as the synthesis proceeds. The buried glucosyl residues thus become inaccessible to @-amylaseand, following exhaustive @-amylolysis,should remain in the P-dextrin. To determine whether the usual mechanism of glycogen synthesis operates in the de nuco system, synthesis was initiated with relatively high specific activity [ 14C]ADPglucose (0.54 mCi/mmole) and after a short period ( 5 min) unlabeled ADP-glucose was added to give a fivefold dilution of the [ 14C]ADP-glucose. The reaction was allowed t o proceed until the amount of glycogen
de Noco
formed was four t o six times that present when the cold ADP-glucose was added. As a control a parallel assay was run in which the chase was carried out with [I4C]ADP-glucose having the same specific activity as that used t o initiate the reaction. After termination of the reactions, the per cent @-amylolysis of the “C in the final product was determined. The data presented in Table VI11 show that, in the absence of glycogen activator, and when BSA is used as an activator, the same fraction of label is cleaved from the molecule whether labeled or unlabeled ADP-glucose is used for the chase. Thus, the per cent of the label that can be cleaved from the product by @-amylase appears to become fixed early in the reaction and to remain unchanged with further synthesis of the molecule. This result is not consistent with that expected for the glucosyl transferasebranching enzyme pathway described above. When the activator is glycogen, it is seen that a significant fraction of the [ 14C]glucoseadded early in the reaction does become buried by the further reaction. Even in this case, however, the glucosyl transferase-branching enzyme mechanism would not predict that such a large fraction of the [14C]glucose would still be accessible to @-amylase after the chase. Several explanations of these preliminary data are possible. For example, the system may initiate the synthesis of many new molecules of glycogen throughout the reaction period rather than continuing to add glucose residues t o those few glycogen molecules formed early in the assay. Alternatively, a pathway similar to that described for the synthesis of branched 0-antigens (Osborn and Weiner, 1968) by successive additions of oligosaccharides to the reducing terminals of the growing molecule (Robbins et al., 1967) would account for
GLYCOGEN
BIOSYNTHESIS
I N
3987
Aerobacter aerogenes
-
BIOCHEMISTRY
GLUCOSYL TRANSFERASE
GLYCOGEN
T
GLUCOSY L
tI
TRANSFERASE (S)
?(&)ADP
BRANCHING ENZYME
BRANCHING ENZYME CHAIN LENGTH TRANSFERASE
T
GLUCOSYL TRANSFERASE
1,
G&L
\
0 1 13
6/13
FiGuw 6: The glucosyl transferase-branching enzyme mechanism for the growth of the A3(S1) glycogen molecule ( F L = 1) in a pulse-labeling experiment. The closed circles represent [14C]glucosylresidues incorporated into glycogen in the early phases of reaction. The open circles are unlabeled glucosyl resides incorporated after the chase. Adjacent circles are joined by 1,4 linkages; the 1,6 linkages at branch points are indicated by arrows. Growth proceeds by succesive transfers of glucosyl residues to the nonreducing terminals by glucosyl transferase (Manners, 1962) until the external chains are sufficiently long to be acted upon by branching enzyme (Larner, 1953; Verhue and Hers, 1966). The action of P-amylase in removal of successive maltose residues from the nonreducing end is shown by the dashed lines. For each structure the fraction of the total number of [14C]glucoseresidues that will be removed by P-amylase is indicated. these data. The latter mechanism, illustrated in Figure 7, is discussed below. Discussion
3988
The purified GS from A . aerogenes catalyzes the synthesis of a highly branched glycogen in the presence of SAP. The absence of glycogen from the purified cell components and from all other components of the assay system has been demonstrated. Thus, the system synthesizes glycogen directly and solely from ADP-glucose. Although a specific heat-labile activator protein has been purified from the A . aerugenes extracts, any one of a group of physically and chemically diverse compounds will also serve as activators. Because of the nonglycogen nature of the compounds which stimulate the activity of GS and their diverse properties, they appear t o serve as activators rather than as glucosyl acceptors. This conclusion is supported in the case of
G A H A N A N D
CONRAD
FIGURE 7: Mechanism proposed for de novo synthesis of glycogen from the reducing end of the molecule (see Robbins et a/., 1967). The glucosyl transferase system is proposed to catalyze formation of the activated “repeating units” of glycogen by transfer of the number of glucosyl residues in E from ADP-glucose to each of the sites (1 and 2) on the enzyme complex. One of the linear a-l,4-linked oligosaccharides thus formed is then divided into E L and E L by the attack of the C-6 hydroxyl group of its potential branching residue on the activated reducing terminal of the linear oligosaccharide attached to the adjacent site on the enzyme. If the attacking residue is always a part of the newly formed E , the glycogen molecule will grow from the reducing end. Note that in the first enzyme
